ES2280372T3 - APPLIANCE AND PROCESS TO SIMULATE AND ANALYZE A SYSTEM WITH FAULT MODES. - Google Patents

Abstract

A process for analyzing a system comprising the following steps: collecting data from a first system, wherein said first system has a plurality of cumulative cause failure modes and a plurality of competitive cause failure modes characterized in that said data they refer to both modes of cumulative cause and competitive cause; parameterize said data for use with a computer program simulating a second system; and executing said computer program simulating said second system, wherein said execution stage comprises the steps of: calculating a first effective working time for each failure mode based on said data collected from said first system, determining which of said failure modes of cumulative cause and failure modes of competitive cause produces a first loss event in said second system by selecting the lowest value of said first effective work times, and if said failure mode that produces said first loss event of said second system it is one of said plurality of cumulative cause failure modes, then calculate a second effective work time only for said cumulative cause failure mode that produces said first loss event for said second system and calculate a second effective work time for each of said plurality of failure modes of competitive cause, and use at least one of said times Possible effective work calculated to implement changes in said system.

Description

Apparatus and process to simulate and analyze a system with failure modes.

Technical Field of the Invention

The present invention relates to the field of processes and devices to analyze a system and, more especially, to the field of processes and devices to simulate and analyze systems repairable, such as manufacturing systems, queuing systems, etc.

Background of the invention

Reliability is a measure of the probability that parts, components, products or systems make their expected functions without failures in specific environments during Desired time periods with a certain level of confidence. Typically, reliability is expressed as a decimal fraction (by example 0.832). Reliability engineering covers the theoretical and practical tools by which it can be specified, predicted, analyzed, demonstrated, installed and Initialized the probability and capacity of which pieces, components, equipment, products and systems perform their functions necessary. The powerful engineering tools of the reliability that allows you to accurately predict the reliability of parts, components and systems can provide the company with a significant competitive advantage. For example, the exact prediction of the reliability of a planned manufacturing or production line or existing can reduce costs, expedite entry into the market with new products and provide results of the most predictable project. The exact prediction of reliability It can also be used to identify and allocate resources for implementation of process changes that can increase the reliability of a manufacturing system, as described in WO-98/24042 A1 or US-5,455,777.

Historically, reliability analyzes and simulations have depended on common statistical assumptions in as for independence and identical distribution of times up to failure event and methods such as chain simulations of Markov (to control state changes sequentially in time) to simulate the reliability of production systems or of its subsystems. These simulations can be slow and faces of developing and executing to simulate the dynamics of a complex manufacturing system. But simplifications can introduce errors that call into question the validity and usefulness of these simulations. The advantages of quick entry into the market, timely entry into the market and about predictable project results, as mentioned previously, they can be obtained thanks to methods that offer a high degree of accuracy in simulating the performance of existing systems and can be used to simulate scenarios that are intended to modify these systems or to simulate systems that can use new combinations of subsystems from valid models of existing manufacturing systems. Costs and manufacturing capacity can be better understood and controlled with methods of greater prediction accuracy giving result in minor market disruptions, especially during First phases of introduction of a new product.

Therefore, there is a need to have simulation methods that can predict the effects of coupling or combination that have premature failures or acceleration and deceleration operations (stop-and-go) on performance of the manufacturing system. In addition, there is a need to have of simulation methods that allow a better use of downtime for repairs and restoration of system operation That is, the repair actions are perform specifically to increase the probability of a Successfully restart the failed subsystem and to reduce the frequency of false starts and short times until failure due to improper repairs (i.e. failures induced by the maintenance).

Summary of the invention

IT devices and processes are provided To analyze a system. The devices and processes incorporate the stages of collecting data from a first system, where the first system has a plurality of cumulative cause failure modes and a plurality of competitive cause failure modes and the data is refer to these failure modes, parameterize the data to use in a computer simulation of a second system and run a Simulation of the second system. The realization of the simulation it also includes the steps of calculating a first effective time of work for each failure mode based on the data collected from the First system, determine which of the cause failure modes accumulated and failure modes of competitive cause produces a first loss event in the second system by selecting the shorter value of the first effective working times, and in case of the failure mode that produces the first loss event of the second system is one of the plurality of failure modes of cumulative cause, then calculate a second effective time of work only for the cumulative cause failure mode that produces the first loss event for the second system and calculate a second effective working time for each of the plurality of failure modes of competitive cause.

Brief description of the drawings

Although the specification concludes with claims that refer in particular and claim from  Clearly the invention, it is believed that the present invention is you will understand better in view of the following description along with the accompanying drawings, where:

Fig. 1 is a schematic illustration of a illustrative process according to an aspect of the present invention;

Fig. 2 is a schematic illustration of a manufacturing system for illustrative paper towels;

Fig. 3 is a schematic illustration of a illustrative hierarchy for the manufacturing system of Fig. 2;

Fig. 4 is a graph of illustrative data of loss of system downtime events;

Fig. 5 is a schematic illustration of Networked computers suitable for use herein invention;

Fig. 6 is a schematic illustration of a preferred architecture for a simulation performed according to the present invention;

Figs. 7 to 9 are schematic illustrations of a preferred process implanted using the simulation of the Fig. 6; Y

Fig. 10 is a schematic illustration of a preferred process implanted for the simulation accumulator of Fig. 6.

Detailed description of the preferred embodiments

Reference will be made in more detail below to the preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings, where equal numbers indicate the same elements in all views. For transparency and simplicity purposes they are used in the Present the following expressions:

Here, the term "system" refers to any set of components, processes, operations or functions that provide a product or service. Illustrative systems include manufacturing systems  and production, airline flight operations, operations maintenance, queuing operations (for example, traffic control, reduction of waiting times on lines or queues such as bank windows, gas stations, boxes supermarkets, etc.), industrial equipment deployments and military (for example, troop deployments, dimensioning of naval aviation groups, rental fleet sizing of cars and trucks or logistics of department stores and of transport).

Here, the expression "event loss "refers to any event that affects negatively to the performance or operation of a system or of one of its components (for example, system stops or times component stop, product quality reductions, increased intervention of inactive operators, trucks or cars, empty lines or queues), where each loss event has a cause and mode of failure associated with it.

Here, the expression "mode of failure "refers to a description of how a system can fail to not perform its intended function. Each failure mode can have one or more causes associated; a stop time during which mode of failure has caused a system or a component of the same stop performing its intended function; and an effective time of work until the next occurrence of a failure mode, time during which a system or a component of it performs its expected function

Here, the term "cause" refers to the reason why a failure mode occurs.

Here, the expression "life" refers to the time elapsed from an initial point to the occurrence of an end point.

Here, the expression "mode of competitive cause failure "(CCFM) refers to a failure mode which is produced according to its own distribution, which is independent of distributions of other failure modes and where time working cash of the competitive cause failure mode ends when any loss event occurs. In the simulations of the present invention, the effective working time of each mode of competitive cause failure is regenerated again after it occurs Any loss event.

Here, the expression "mode of cumulative cause failure "(CMFM) refers to a failure mode which is produced according to its own distribution, which is independent of distributions of other failure modes and where time Work cash from the cumulative cause failure mode is not visible affected by the occurrence of unrelated loss events with the cumulative cause failure mode in question. In the simulations of the present invention, the effective working time each mode of cumulative cause failure is only regenerated again upon expiration

Here, the expression "event false start "refers to a loss event that produces quickly with respect to the expected life of a system (for example, two minutes or less for a system of manufacturing that can have an expected half-life of twenty to thirty minutes) after a system, or a component of the It has restored acceptable operation.

Here, the expression "system in series "refers to a system that fails if any fails of its components.

Here, the expression "system in parallel "refers to a system that fails if all fail its parallel components.

Here, the expression "time work cash "refers to the period of time or life of a system, or component thereof, during which it performs its expected function

Here, the expression "time Stop "refers to the period of time or life of a system, or component thereof, during which it does not perform its function expected due to a loss event.

Here, the expression "availability" refers to the relationship between total time working cash for a system and the sum of total time working cash of a system plus the total stop time of a system.

Here, the expression "time mean between failures "(MTBF) refers to the relationship between the Total effective working time of a system and the total number of System loss events.

Here, the expression "time means of repair "(MTTR) refers to the relationship between the total stop time of a system and the total number of events of system loss

Here, the expression "parameterize" refers to the process of identifying or adjusting data according to a parametric equation (for example, an equation that Contains parameters of shape, scale, and location to describe data, such as Weibull equations, equations log-normal, normal equations, etc.).

Referring to Fig. 1, and according to a aspect of the present invention, a process to analyze and / or simulate an illustrative system. The process 20 begins in stage 21 where a hierarchical level is selected of the system. The hierarchical level defines the desired system analyze and represent the simulation depth of the components of a system. For ease of discussion, the The present invention will be described herein with respect to to a system in the form of a manufacturing system and machines that are part of the manufacturing system. More especially, the The present invention will be described herein with with respect to a manufacturing system 22 of wipes in series, such as It is schematically illustrated in Fig. 2. The system of manufacturing 22 comprises a sheet roller 24 which has rolled around it a sheet of paper 26 with a finite length. The sheet 26 passes through an embossed printer 28 that prints a pattern on sheet 26. Sheet 26 is cut in the direction of the machine in discrete parts 32 in a first cutter 30. The parts 32 are wrapped around a cylindrical paper core in the core roller 34 after which the rolled parts 36 they are deposited in an accumulator 38. The rolled parts 36 are then cut in the transverse direction to the machine in the second cutter 40. The rolled parts 42 cut twice are supplied to a wrapper 44 that wraps a plurality of rolled parts 42 in a polymeric film 46 which is extracted of a roll 48 of polymeric film. A plurality of sensors 50 are distributed throughout the manufacturing system 22 to detect events of loss of the manufacturing system, as will discuss in more detail later. Although the system Manufacturing 22 is described herein at a level Machine hierarchical, it will be appreciated that the manufacturing system 22 can be described at higher or lower hierarchical levels according to the analysis in question and the level of simulation accuracy desired and that these different levels can be simulated from jointly or individually, as desired. Fig. 3 presents different illustrative hierarchies superior and inferior to the described for the manufacturing system 22 of Fig. 2. By For example, a higher hierarchical level may include a network of distribution 52 comprising a plurality of plants of manufacturing 54 (a parallel system), where each has a plurality of manufacturing systems 22, wherein the system of manufacturing 22 has a plurality of machines 56 (for example, the roller 24) associated therewith and each machine 56 has a plurality of subcomponents 58 (for example, a bearing for the sheet roller 24). In addition, although the present invention is described herein with respect to a system of manufacturing, it will be easily appreciated that also other products, services, manufacturing systems and systems are adequate for use in the present invention, as described previously.

Referring again to Fig. 1, analyze the system hierarchy selected in step 21 to identify their accumulated and competitive failure modes as well as its causes A list is presented in Table 1 below. illustrative of cumulative and competitive failure modes and their causes for the manufacturing system 22.

TABLE 1

one

       \ vskip1.000000 \ baselineskip
    

By way of illustration with respect to Table 1, the sheet roller 24 has three failure modes: two modes of cumulative cause failure and a competitive cause failure mode. He the fact that a failure mode is accumulated or competitive can be based on the cause of the failure mode or the rules of the structure (for example, all failure modes are competitive unless otherwise indicated). Structure rules can be adapted according to the objectives or uses of the simulation. The mode of leaf jam failure is competitive because instability of the leaves can re-produce a leaf jam after Having repaired any loss event. The failure modes of belt and roller change (i.e. roller replacement because the sheet is sold out) they are accumulated because they depend on the finite life of the belt and blade, regardless of others loss events not related. Failure modes may have multiple causes (see, for example, failure mode number 1 in the Table 1) and stop time distributions. In addition, the modes of failure do not necessarily present a correspondence between the number of causes of failure mode and the number of distributions of stop times Distributions of stop times represent independent features with a distribution Identical repair times for a single failure mode. By example, the leaf jam failure mode has three distributions of downtimes, namely a short period of repair time, an average period of repair time and a Long period of repair time. The selection of the number of downtime distributions can be made using methods known in the art, such as segmented regression, in where each discontinuity between segments represents a separate distribution The probability C of falling within any distribution of determined downtimes is determined by the relationship between the number of loss events within a distribution of downtimes and the total number of events of loss to the failure mode in question. For example, in Fig. 4 an example of a graph for cause failure mode is presented Competitive jamming of sheet roller 24. The number Total loss events illustrated in Fig. 4 is 188. For the first distribution of stop times 70, the number of loss events is 109 and therefore the probability C that produce a stop time in the first time distribution Stop 70 is 0.58. Similarly, the number of events of loss for the second distribution of stop times 73 is of 76 and, therefore, the probability C of a time of stop in the second distribution of stop times is 0.40.

In relation again to Fig. 1 and the process 20, the system loss event data based on the time (for example, effective working times, times of stop) are collected in manufacturing system 22 and analyzed parametrically in stages 60 and 62 once the level is selected hierarchical manufacturing system 22. The collection of these System loss event data is used to characterize  and quantify the failure modes of the manufacturing system 22. The System loss event data collected preferably include causes of failure mode, time stamps that mark the beginning and end of periods of effective working time and of stop time, and the absolute time of effective times of work and stop times. Depending on the cause of the failure mode, different detection techniques can be used To collect the data. Table 2 below presents Method examples

TABLE 2

3

As indicated, the event data of System loss can be picked up using 50 sensors and / or programmable logic controllers (or PLCs). Examples of System loss event data for the system Manufacturing 22 are presented in Table 3. A sufficient data volume for each failure mode in order to Establish a statistical sample. More preferably, at least Approximately 3 events are recorded for each failure mode. Most preferably, between approximately 15 events and Approximately 30 events are recorded for each failure mode. System loss event data may also include other information, such as a date stamp. Although Table 3 is refers to system loss event data for a system manufacturing, other types of data can be controlled in others systems such as idle time in the case of queue systems (for example, gas station available but inactive).

TABLE 3

4

Once the event data has been collected from system loss, these are parameterized in step 62 of process 20. The data of the system loss event can be parameterized using one of the different methods of distribution / statistical function known in the art, such as exponential distribution, normal distribution, distribution of Weibull or normal logarithmic distribution. The distribution of Weibull (in particular the probabilistic density function of Weibull or PDF) is especially useful for identifying times work force while stop times are typically determined by a normal distribution Logarithmic False boot events are preferably calculated using the following equation:

1 - (total number of events of false start for a failure mode) Probability false starter (P) = -------------------------------------------------- ------------ (one) Number total events of lost for him system

       \ vskip1.000000 \ baselineskip
    

Therefore, a probability of false start-up P of 0.97 (or 97%) means that 97% of the time there is no false boot for this failure mode after any event of loss for any mode of failure. A typical Weibull PDF for The effective working time has the following form:

Probability of failure = (Beta / alpha) (time work cash / alpha) ^ (Beta-1)} e ^ {(- (((\ text {time cash of work / alpha) Beta})))} (2)

where the constant alpha and beta are the usual Weibull constants (i.e. alpha is the scale parameter and beta is the shape parameter) known in The technique. Equation (2), when solved for time working cash, can be reorganized in the way next:

(3) Time working cash = alpha (-ln (probability of failure)) ^ (1 / Beta)}

Equation (2) can be used to calculate effective working times based on the generation of a random probability using a uniform random real number between zero and one for the probability of failure. Can be used any log-normal equation known in the art that has sigma parameters (that is, the standard deviation of the logarithm of stop times) and mu (that is, the average logarithmic stop times) to parameterize a distribution of downtimes. Only for discussion herein, the effective time distributions of work for manufacturing system 22 failure modes, presented in Table 1, are parameterized using a PDF of Weibull and stop times are characterized using a log-normal function, where the constant C represents the probabilities between the different distributions of stop times. It should be noted that the sum of the values of the constant C of any failure mode must be equal to one because the stop time of any loss event must enter within one of the stop time distributions of the present invention for the failure mode of this loss event. The probability of false start of each failure mode is represented by a constant P. In Table 4 below, present examples of alpha, beta, mu, sigma and C and P constants for Each mode of failure. As mentioned earlier, some of Failure modes have more than one time distribution of stop. For example, a belt-activated roller may fail to cause of a drive belt failure. The roller can be quickly repaired if multiple straps are hung around of the drive pulleys so that it is not necessary disassemble the structure of the roller for repair, reducing So the repair time. However, once it has been used the last repair belt, it would be necessary to disassemble the structure of the roller, which would lead to greater distribution of repair time than if there were a strap has been available for repair. They are also useful multiple distributions of stop times to model modes of failure that have more than one cause, each cause having a repair distribution of separate stop. The number of stop time distributions can be selected using one of the different methods known in the art (for example, segmented regression).

TABLE 4

5

6

A constant is also included in Table 4 of speed R representing a product speed value in steady state for a simulation object (for example, 54.9 m / min [180 ies per minute] roller sheet feed of leaves).

After completing the collection and parameterization of System loss event data in stages 60 and 62 of process 20 of Fig. 1, a computer program is designed that models or simulates manufacturing system 22 according to another aspect of the present invention. The simulation program is run or run on any special purpose computer or general or other digital processing apparatus, such as a desktop computer, a server and / or client computer united in network (for example, over the Internet or an Intranet), a micro-computer, manual organizers or others forms of computers and computer systems otherwise known in the art. Examples of networked computers 1000 are schematically illustrated in Fig. 5. The computer preferably it comprises a logic circuit (such as a unit of  1030 central processing, a microprocessor or other microcontroller) capable of running the simulation program. He simulation program, or parts thereof, can be provided as a program product, where the program product includes a signal carrier medium that can be configured to store data and / or machine-readable instructions that make that the logic circuit to which it is attached perform the steps of Simulation discussed below. The signal carrier medium can be provided in the form of an optical disk, a magnetic disk, a magnetic hard disk drive (for example, number of reference 1010), a magnetic tape, RAM, ROM, or any other magnetic, optical or other storage media readable by computer. Alternatively, the program product may be distributed with instructions contained in other media signal carriers, including digital communication links and analogs (for example, such as a part of wire or fiber of a local area network, a part of wire or fiber of a network of wide area, a part of a wireless network, etc.), a wave carrier or propagated signal and other forms of media transmission. The computer also preferably comprises one or more peripheral input / output such as keyboard, mouse, touch screen, microphone, monitor, printer, etc., which can be connected to the logic circuit through a system bus and a adapter (for example, display adapter 1040). Although the step 148 of Fig. 1 is preferably computerized, it is you will appreciate that the other stages (or parts thereof) of the process 20 can be implanted by or together with an apparatus of digital processing

In Figure 6, a simulation program 80 preferred comprises a plurality of simulation objects 82 and a controller 84, wherein each simulation object 82 represents a physical component of the modeling system (e.g., a plant, a manufacturing line, a machine, a component, a glue, etc.). The simulation program can be implemented using any of a number of programming languages or not oriented to objects known in the art (e.g., C, C ++, EXCEL macros, etc.). In this case, the simulation objects each one represents a machine 56 of the manufacturing system 22 and comprise instructions and / or data describing the operation of said machine. Therefore, there are illustrated simulation objects in Fig. 6 for the sheet roller 24, the embossed printer 28, the first cutter 30, the core roller 34, the accumulator 38, the second cutter 38, the roller 42 of polymeric film and the envelope 44. The simulation objects 82 are connected with their corresponding upstream and current simulation objects below so that a representation is arranged virtual manufacturing system 22. Simulation objects 82 preferably they send their simulation objects immediately adjacent upstream and downstream simulated speed (for example, a linear speed in feet / minute of sheet 26 between simulation object 1 and simulation object 2) and the simulated state of the simulation object (for example, stopped or operational). The controller 84 coordinates the simulation time (a continuation "current simulation time") between different simulation objects 82. Preferably, the coordination is performed for each simulation object at transmit your effective work time and / or downtime to controller 84 when each simulation object finds an event simulated loss (for example, one stop). When it is found a simulated loss event, controller 84 advances the time of Current simulation according to the stop time or the effective time of work calculated on a simulation object, as described in More detail later.

A process is described in Figures 6 to 8 illustrative 86 implanted together with a simulation object 82 of according to another aspect of the present invention. The stages of process are described both generically for any object of simulation 82 as an illustration for the simulation object of the sheet roller, although it will be appreciated that these stages can be implanted in a similar way for the other objects of simulation 82 of simulation 80. Although the stages of the process 86 are described herein with respect to objects of simulation, some stages implemented by controller 84 also They are incorporated into process 86 to facilitate discussion. Too it will be appreciated that the stages of process 86 and the disposition of Simulation and controller objects can be modified, reorganized, combined and separated as known in the art without abandoning the scope of this invention.

Process 86 begins in step 88 where the parameters (for example, constants P / C / R, alpha, beta, mu, sigma, etc). for distributions of effective working time and of downtimes, speeds and probabilities of false start of simulation object 82 are preferably read in a file input or similar object. Once each object of simulation 82 has been thus initialized, a series of probabilities of false start in stage 89, taking preferably the series the structure presented in Table 5 next. The first column is the constant P of probability of false boot while the second column is the difference of one minus the probability of false start P. The third column is  the quotient of one minus the constant P of probability of false start between the sum of one minus the probability constant P of false start of column 3. The fourth column is the sum Current of the third column.

TABLE 5

7

Then a probability of false is added target start for competitive cause failure modes of the Simulation object according to the following equation in step 90:

(4) Probability of false target start = 1- \ Sigma (1- probabilities of false start P)

After determining the probability of false target start in stage 90, a real number is generated uniform random between zero and one in step 94 and this number uniform random real is used to calculate an effective time working for each cause failure mode accumulated for the object of the present simulation 82 using equation (2) above. For example, the sheet roller 24 has the failure modes of cumulative cause of roller change and belt failure, as Indicate in Tables 1 and 4. Illustrative alpha and beta values for failure mode number 2 indicated in Table 4 are 63.00 and 15, respectively. For a uniform random real value of 0.54, this mode of cumulative cause failure would have an effective time of 63.46 minutes work in stage 94. These effective times of work are preferably stored in one series or another data structure that is updated when the simulation 80.

Stages 96 to 106 are below. executed to calculate a false start error from failure mode  of competitive cause. In step 96, an event counter of loss that stores the total number of loss events simulated (i.e., failure events of failure modes of both competitive cause as of accumulated cause) is increased each time that a loss event is found. If the previous event of loss to the simulation object was due to a mode of competitive cause failure and a false start had occurred as indicated in step 98, step 100 is executed, increasing by one the counter that adds the number of false Competitive cause failure mode starts. Otherwise, it execute step 102 directly, calculating the probability of simulation false start events using the following equation:

(5) Simulation of prob. false start = \ frac {\ text {Total number of events of CCFM false boot}} {\ text {Total number of loss events of the simulation}}

The total number of loss events of the simulation in the denominator of equation (5) has already been calculated  in step 96, and the total number of false boot events of competitive cause failure mode in the numerator of the equation (5) has already been calculated in step 100. Next it is calculated in step 106 the error between the probability of false start of the simulation and the probability of false target start according to the following equation:

Error = probability of false start of the simulation - probability of false start objective

Once the error has been calculated, it calculates in stage 106 a probability of false start of the Simulation corrected using the calculated error. Can be used any error correction method known in the art to calculate the corrected value of the probability of false simulation start (for example, a binary search). Although it is described herein by way of simplification  an error correction regarding false starts after a competitive cause failure mode, it will be appreciated that also a similar error correction could be implemented for fake starts after an accumulated cause failure mode.

In step 110 it is determined whether the above loss event was due to a cause failure mode competitive or if it was due to a cause failure mode accumulated If it was due to a competitive cause failure mode, steps 112 and 114 are used to determine if there is a false start after stop time (i.e. after time repair) for the previous loss event considering the probability of false start of the corrected simulation determined in step 106. If the previous loss event has been due to a cumulative cause failure mode, the stages are executed 111 and 113 (Fig. 8). In step 111 a real number is generated uniform random between zero and one. If the random real number uniform is less than or equal to the probability of false starting P of the previous mode of cumulative cause failure, as described in step 113, then a false start has occurred after This loss event. Then the time of stopping this false start event using process 116 (Fig. 9), as discussed in more detail below; of what Otherwise there is no false start of accumulated cause after of the previous cumulative cause loss event and run at then step 112 to determine if there is an event of false start of competitive cause that begins in stage 112, as is also discussed in more detail below.

In step 110, if the previous stop was a event of loss of competitive cause (or if stage 110 is run for the first time when the simulation starts), the stages 112 and 114. A uniform random real number is generated between zero and one in step 112 and if this uniform random real number is greater than or equal to the probability of false starting of the corrected simulation determined in step 106, so it has not been produced a false start after the previous loss event and parallel routes A and B are performed (Fig. 8) to determine the following minimum effective working time for the purpose of the present simulation 82. If the uniform random real number generated in step 114 is greater than the probability of false corrected start determined in step 106 (i.e. there is a false start event after the previous loss event), to a new uniform random real number is then generated in the step 115 and this number is compared with the fourth column of the series generated in step 89 to determine which failure mode of the machine stage of the present invention has produced the false start. For example, with reference to Table 5, if the number Uniform random real generated in step 115 is 0.85, then failure mode number 3 would be used to determine the time of stop for the false start event because 0.85 is greater than 0.7777 but less than 1. After determining which mode of failure the false boot event occurs, the process of stop 116 (Fig. 9) to determine the stop time for the false boot event.

In Figure 8, the parallel route A calculates the effective work times for each cause failure mode competitive object of this simulation while the parallel route B generates the effective work times for each mode of failure of cumulative cause of the object of the present simulation. Starting with step 118 for route A, it is generated separately a uniform random real number between zero and one for  each competitive cause failure mode of the simulation object. Each uniform random real number is applied to equation (3), in where equation (3) incorporates the alpha and beta values for the mode of failure corresponding to the uniform random real number generated. For example, Table 6 presents the random real number uniform generated for each competitive cause failure mode of the sheet roller 24. Using the alpha and beta values for each failure mode, an effective working time is calculated based on the uniform random real number generated for a failure mode of competitive cause Then in step 120 the Simulation time of lowest competitive cause failure mode of all cause failure mode simulation times competitive for the purpose of this simulation and this value is sent to stage 132.

TABLE 6

8

Referring now to parallel route B of Fig. 8, the time is determined in steps 122 to 130 lower working cash between cause failure modes accumulated for the purpose of this simulation 82. In the stage 122, for each mode of cumulative cause failure of the object of the This simulation 82 subtracts the effective working time of the effective working time system in cause failure mode accumulated from each mode of cumulative cause failure and these new values are stored in the series. In step 124 it is determined whether The cause of the previous loss event is a cumulative cause. In positive case (for example, a cumulative cause failure mode has produced a stop), then a new effective time is calculated working in cumulative cause failure mode for this mode of cumulative cause failure generating a random real number uniform between zero and one in step 126; otherwise, it execute step 130 as described below using the equation (3) and the alpha and beta values for the failure mode of cumulative cause of the present invention, the new effective working time that replaces the effective time of work in cumulative cause failure mode expired in the series. TO then in step 130 the lowest percentage is selected of effective working time in cumulative cause failure mode of  among all the percentages of effective working time in mode of cumulative cause failure available in the series for the object of the present simulation 82. Next, in step 132, select the lowest effective work time between the stage 130 and step 120 to determine which cause failure mode competitive or what mode of cumulative cause failure will produce the following loss event for the purpose of this simulation 82.

After determining the failure mode produced by the next loss event for the simulation object, this lowest percentage of effective work time in failure mode is sent to controller 84 in step 133. In step 135, the controller 84 selects the lowest percentage of effective time of work in failure mode among all objects of simulation (the "effective working time of the system"), including loss events of any product level of accumulator (discussed in more detail below as another preferred embodiment) and add this effective working time to the current simulation time, thus advancing the simulation at Loss event start. The effective working time of system is returned to each simulation stage. In step 139, each simulation object determines whether its effective time of work in failure mode determined in step 132 matches the Effective system work time returned. In case Yes, the execution goes to the stop process 116 which performed beginning at step 134 (that is, this object of simulation is the one that finds the loss event), as shown in Fig. 9; otherwise, steps 141 and 143 where the simulation object enters a state of "wait or inactivity" in step 141 and each object of simulation awaits the entry of its simulation objects adjacent and / or controller. In relation to Figure 9 and the stop process 116, updated speed information and status is sent to each adjacent simulation object in the stage 134, which is usually a speed for a loss event equal to zero and a status indicator that the object of the Present simulation 82 has stopped. Status indicators only they are used to coordinate events between the objects of simulation. In step 136 a random real number is generated uniform between zero and one and this number is used in the stage 138 to select the distribution of stop times to be you want to use to calculate the stop time for the event of loss in question comparing each constant C of stop time with the uniform random real number. The constants C are added sequentially and the uniform random real number is compared to the interval between the sums to determine the distribution of stop times to use. For example with reference to Table 4, failure mode 1 and Table 7, if the uniform random real number generated in step 136 is 0.85, then to determine the stop time the distribution of stop times 2 in Table 4 because 0.85 is greater than 0.6 but less than 0.95.

TABLE 7

9

       \ newpage
    

In step 140 a new real number is generated uniform random between zero and one. This random real number uniform is used in equation (4) to determine the time of stop for the loss event in step 142 using the distribution parameters of stop times for distribution of stop times determined in step 136. The time of stop calculated in step 142 (the system stop time) is sent to controller 84 in step 144. The controller adds the system downtime to the current simulation time (it is that is, the simulation time is now advanced at the end of repair period) and send the current simulation time to each of the simulation objects in step 144 (this information is received by simulation objects in the stage 143). The speed and status for the purpose of this simulation 82 that has produced the loss event are after updated and sent to adjacent simulation objects in step 146, where the speed is generally set to the speed constant (for example, 54.9 m / min [180 feet / minute] for sheet roller 24) and the status is updated. The execution, for all simulation objects 82 is reset to then at point E of Fig. 6, beginning the correction of the probability of false start error in step 96.

A preferred alternative embodiment of the The present invention incorporates the process 152 illustrated in Fig. 10 for simulation objects that model storage components or buffer, such as accumulator 38. Each accumulator has an empty product level value (that is, the value in the that there is no product in the accumulator), a level value of current product (that is, the current value of the amount of product in the accumulator) and a full product level value (it is say, the value at which the accumulator is completely full with product and cannot accept more product). Starting at the step 154, the difference between the product speed is calculated that the accumulator enters and the product speed that leaves the accumulator. By way of illustration with respect to simulation 80, the product speed entering the accumulator 38 of the manufacturing 22 is the difference between the speed constant for the product supplied from the simulation object 82 that represents the core roller 82 and the speed constant for  the product received by the simulation object 82 that represents to the second cutter 40. If the difference is greater than zero in the step 156, then step 158 is executed, determining the time left before the accumulator 38 is full by subtracting the product level value full of product level value current and dividing this difference between the speed difference calculated in step 154. This value is then added to time. of current simulation received by simulation object 82 for the accumulator in step 143 of Fig. 6 and sent to step 132. This accumulator level simulation time value (which represents a loss event) is then compared to competitive cause and cumulative cause simulation times more lows of stage 132. With respect to stage 160, if the difference speed is less than zero, then step 162 is executed, where the time left for the accumulator 38 to empty is determined by subtracting the current product level value from the value of empty product level and dividing this difference by speed difference calculated in step 154. This value is then added to the current simulation time received by the simulation object 82 for the accumulator in step 143 of the Fig. 6 and sent to step 132. This simulation time value Accumulator level (representing a loss event) is then compared to competitive cause simulation times and of cumulative cause lower than stage 132. Otherwise, it perform step 164 (i.e. the speed difference is the same to zero and the input speed is equal to the output speed), where the level of product inside the accumulator is constant and an accumulator level simulation time value equal to Infinity is sent to stage 132.

With respect to Table 8, below is describe six illustrative time stages (for example, T = 0 minutes, T = 143 minutes, T = 322 minutes, T = 327 minutes, T = 330 minutes and T = 350 minutes) for a simplified simulation according to the present invention of the manufacturing system 22. Each stage of time incorporates one or more stages of simulation processes described above and aims to show in a general way and for greater clarity of the concepts described herein. In the first stage of time (i.e., T = 0 minutes or the beginning of the simulation after initializing the simulation) a number is generated uniform random real in the fourth column to determine if there is a false start for cause failure modes competitive under the simplified assumption that there is no false cumulative cause start when starting. The same only real number Random is used for each competitive cause failure mode of an object of the machine of the present invention (i.e. steps 112 and 114 of Fig. 7) and a random real number is generated separated for each cumulative cause failure mode when it results appropriate (that is, when step 111 of Fig. 7 is executed). If there is a false start as determined in the stages 113 or 114 of Fig. 7, a "yes" is introduced in the fifth column. Since there is no false start in the first stage of  time, a uniform random real number is generated for each mode of failure and this uniform random real number is then used to generate an effective working time. In the eighth column you select the lowest effective working time for each object of simulation and then in the ninth column the lowest effective working time for the system. Time System work cash is added to simulation time current in the second stage of time (that is, time T = 0 + 143 minutes) A uniform random real number is generated in the tenth column though, because there is only one time distribution  stop this does not affect the distribution of stop times selected In the eleventh column a random number is generated which is then used to determine the stop time for the failure mode number five. In the third stage of time, the stop time of the twelfth column is added to the time of simulation (that is, time T = 143 + 179). Real numbers random uniforms are again generated in the fourth column for the probabilities of false starting (except the failure modes of accumulated cause since the previous loss event was not due to an accumulated cause failure mode) and there are no false starts. New effective working times are generated for the competitive cause failure modes only in the seventh column based on the uniform random real numbers indicated in the sixth column The effective working times for the modes of cumulative cause failure in the seventh column are equal to time working cash of the first stage of time (i.e. time T = 0) minus the effective working time of the system calculated in The first stage of time. A new effective time is selected system work in the tenth column (for example, 5 minutes) After the calculation of the fourth stop time time stage (i.e. time T = 327 minutes), a false start in the fifth stage of time (that is, time T = 330 minutes) and a stop time is calculated for the failure mode eight that produced the false boot event. This is calculated according to steps 134 to 146 of Fig. 9. In the sixth stage of time (i.e. time T = 350 minutes), the stop time for the false start event calculated in the fifth stage of time is add to the current simulation time in the first column of the fifth time The effective working time for each mode of competitive cause failure is generated in the seventh column using the uniform random real number of the sixth column. Since there has been no effective working time of the system between the fourth and sixth time stages, the times work force for cumulative cause failure modes they remain the same as the effective working times of the Third stage of time.

       \ vskip1.000000 \ baselineskip
    

       \ vskip1.000000 \ baselineskip
    

       \ vskip1.000000 \ baselineskip
    

(Table goes to page next)

10

eleven

12

13

Referring again to Fig. 1, after setting up simulation 80 in step 148 as has described above, simulation 80 can be used to analyze manufacturing system 22 and implement changes in the manufacturing system 22 to improve its reliability, as described in step 150. For example, the size of the accumulator 38 can be modified, the speed constants for a machine can be adjusted, the series of machines in the system manufacturing can be modified, the value to eliminate a mode of fault can be modified, etc. In addition, system data can be used to model a second, not yet built, manufacturing system that has similar system objects but perhaps arranged differently to predict and optimize the reliability of the second system before building it.

Below are examples of availability, MTBF and MTTR for an existing manufacturing system  and the same values of a simulation performed according to the present invention, wherein the simulation models failure modes accumulated and competitive with beta values greater and less than one and probabilities of false start. Preferably, the percentage error (i.e., the relationship between the actual value minus the value simulated and the actual value X 100) is less than about 3%, and more preferably less than about 2% and with maximum Preference the error is less than about 1%. How I know you will appreciate, in a simulation carried out in accordance with this invention other values such as the number of loss events or downtime per loss event. The magnitude of the error can be further reduced, for example, by increasing the number of simulated stop time distributions.

Example 1

14

Example 2

fifteen

Example 3

16

Claims (9)

1. A process to analyze a system that It comprises the following stages: collecting data from a first system, wherein said first system has a plurality of cumulative cause failure modes and a plurality of competitive cause failure modes and characterized in that said data refers to both modes of cumulative cause failure and of competitive cause; parameterize said data for use with a computer program simulating a second system; Y run said computer program simulating said second system, wherein said execution stage comprises the stages of: calculating a first effective working time for each failure mode based on said data collected from said first system, determine which of these modes of cumulative cause failure and competitive cause failure modes produces a first event of loss in said second system by selecting the lowest value of said first effective working times, and if said mode of failure produced by said first loss event of said second system is one of said plurality of cause failure modes accumulated, then calculate a second effective working time only for said mode of cumulative cause failure that produces said first loss event for said second system and calculate a second effective working time for each of said plurality of competitive cause failure modes, and use at least one of said effective work times calculated to implement changes in that system. 2. The process of claim 1, characterized in that said execution step further comprises the step of calculating a stop time for said failure mode that produces said first loss event. 3. The process of claims 1 and 2, characterized in that said execution stage further comprises the step of determining whether said second system will find a first false start event in view of the data collected from said first system. 4. The process of claims 1, 2 and 3, wherein said execution stage also includes the following stages: if there is a false start event, then calculate a stop time for said first false start event; characterized by the stage of determine if said second system will find a second false start event after that time of stop for said first false start event. 5. The process of claims 1, 2, 3, and 4, wherein said execution stage further comprises the step of providing a reliability value for said second system, characterized in that the error of said reliability value is less than approximately three percent. 6. A process to analyze a system, which It comprises the following stages: receive values for a plurality of times work force and downtime for a first system, in wherein said first system has a plurality of failure modes of cumulative cause and a plurality of cause failure modes competitive, referring to these values for a plurality of effective work times and stop times at both modes of failure of cumulative cause and competitive cause and calculate a first effective working time for each failure mode based on said data collected from said first system, determine which of said modes of cumulative cause and competitive cause produces a first event of loss of said second system by selecting the highest value under said first effective working times, characterized in that if the failure mode produced by said first loss event of said second system is one of said plurality of accumulated cause failure modes, then calculate a second effective time of work only for said cumulative cause failure mode that produces said first loss event for said second system and calculate a second effective work time for each of said plurality of competitive cause failure modes, and use at least one of those times workforce calculated to implement changes in said system. 7. A program product comprising a signal carrier medium carrying out an instruction program machine readable by a processing device digital to perform the steps of claim 6. 8. The program product of claim 7, wherein said signal carrier means is at least a part of a computer network

         \ newpage
      

9. A manufactured item comprising: at least one computer; Y a program product comprising a medium signal carrier that performs a readable instruction program by machine executable by a digital processing apparatus for performing the steps of claim 6.